Should notebooks be executed before they are uploaded to AE? about examples HOT 4 CLOSED

I'm not sure what you mean. If it's something to be run with panel serve, the notebook outputs will be ignored; it essentially just runs the .ipynb as .py, where outputs are irrelevant. Or are you asking about a different scenario?

from examples.

I am asking about deploying notebooks as notebooks not as apps. This is for the case where the example isn't panel-related. For example see https://attractors-python36.aip.anaconda.com/

from examples.

Ah, I see. In most cases the examples don't need to deploy at all; they can simply be static web pages as part of a pyviz-topics website. That attractors example is a special case of mixing interactive code with lots of static text and images, which is probably a bad idea -- instead we should probably pull out the interactive bits as a separate app in the same project directory, to make it clear what's static and what's not.

In any case yes, in the (presumably rare) case when we want to deploy a notebook as a notebook, then I agree it probably needs output included before the user sees it.

from examples.

Ok so the consensus is that we should not check in or upload evaluated notebooks. Most of the time we won't be deploying notebooks anyways, but when we do, it would be ideal if we could start the notebook kernel and run the notebook when the user enters the notebook. This would be implemented as a new argument in anaconda-project. anaconda/anaconda-project#198

from examples.